INDEX

Explanations

expressions related to manners, specifically focusing on concepts of rudeness and politeness

oai_token-act-pair · gpt-3.5-turbo

terms related to rudeness and politeness

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lisher

-0.86

ernels

-0.83

ishop

-0.75

panel

-0.75

hart

-0.74

razil

-0.73

ARK

-0.71

lished

-0.68

ilation

-0.68

yrinth

-0.68

POSITIVE LOGITS

 rude

0.93

rud

0.87

 etiquette

0.87

 awakening

0.81

 polite

0.81

 manners

0.79

 disrespectful

0.78

 disrespect

0.78

 greeting

0.76

 respectfully

0.73

Activations Density 0.041%

expressions related to manners, specifically focusing on concepts of rudeness and politeness

terms related to rudeness and politeness

No Comments

No Known Activations